The software testing market has undergone a major shift over the last few years, and we believe that it is on the cusp of another shift. In the past, software testing was primarily conducted by a team of Quality Assurance (QA) professionals who manually verified whether the application functioned as intended. As release times shortened, companies adopted more agile testing processes and instead of relying on manual testing, have shifted to automated testing.
Automated software testing is the process by which dev teams test their own code by writing codified logical and functional tests. These tests run faster than a manual tester could test code. They can also be run frequently under different conditions and a changing code base.
The consequence of this shift in the testing processes is that the responsibility of code quality now lies with the developers themselves and no longer with QA. Some estimate that developers spend as much time on writing their testing code and verifying code logic as they do writing their primary code. Not to mention the fact, that testing code is considered a chore. As a result, tests are often poorly written, not documented nor truly improve the quality of the code. This is why even the very best applications still suffer from downtime, bugs in production and security vulnerabilities.
This problem deepens when you add that developer turnover is high and when they leave it is near impossible to understand the code they developed. As well as the proliferation of Open Source code, which often isn’t understood well by development teams.
Given this context, we are excited to announce our investment in CodiumAI, a code integrity platform that enables developers to maintain the integrity of their rapidly scaling code bases and create high-quality code that works as expected. The company has raised a $10.6 million seed round co-led by TLV Partners and Vine Ventures.
CodiumAI’s platform requires both deep AI expertise and developer tool knowledge. Co-Founders Itamar Friedman (CEO) who served as Alibaba Group’s Director in their Israeli AI Lab, and Dedy Kredo (CPO) who led product and data science teams at Explorium and VMware are just that. Their unique combination of deep AI domain knowledge and being developers themselves is crucial for creating such a solution.
As code is written, CodiumAI concurrently generates tests and checks to ensure the integrity of the new code, CodiumAI analyzes the application’s code and docs, leverages state-of-the-art AI, and interacts with the developer, to help generate documentation and tests that demonstrate the existing code behavior. It takes a load off developers’ shoulders giving them the freedom to focus on their primary code.
While still in its infancy, code generation tools are showing promise. Code generation, a category within the broader Generative AI space, has experienced some extraordinary advances in its underlying technology. In particular, the advances made by OpenAI and its Codex algorithm that translates free text into code. Github’s Copilot, which is a tool driven by Codex, proved the value of code generation tools, with 1.2 million developers registering for the product in 2022. In fact, it is my belief that we are only getting started with code generation and its impact will cause tidal waves. Or as my colleague Shahar Tzafrir, Managing Partner at TLV Partners, put it elegantly “A new way will replace the practice of programming software. The old way was to program code by highly trained, exceptionally bright individuals. The new course will be for code to write code.”
CodiumAI works alongside developers. It enhances the developers’ ability to create high-quality code. It also interfaces with the developer when he/she feels most natural within the IDE.
What is for certain is that Code Generation tools will multiply the amount of code being developed and generated. This will make it even harder for developers to ensure code quality. We are scaling code beyond what automated testing will be able to cope with. Tools like CodiumAI are therefore necessary to keep up.
We are proud to support CodiumAI in their mission to simplify code integrity and wish them the best of luck on this journey. We are excited to be a part of it.
